StackML/benchmark.py at master · ggiesa/StackML · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
from pmlb import fetch_data
from sklearn.linear_model import LinearRegression, LogisticRegression
from sklearn.ensemble import RandomForestRegressor, RandomForestClassifier
from sklearn.neural_network import MLPRegressor, MLPClassifier
from sklearn.svm import SVR, SVC

from stackml import StackML

from importlib import reload
import stackml as s
s = reload(s)


def get_benchmark_data(dataset_name, split_percent=.9):
    data = fetch_data(dataset_name)
    y = data.target
    X = data.drop('target', axis=1)
    split = int(len(X)*.9)
    return X.iloc[:split,:], y.iloc[:split], X.iloc[split:,:], y.iloc[split:]


def classification_benchmark(StackML_model=None, datasets = None, models=None):

    if not datasets:
        datasets = [
            'breast-cancer',
            'cleveland-nominal',
            'horse-colic',
            'solar-flare_1'
            ]

    if not models:
        models = {
            'Logistic Regression: ':LogisticRegression(),
            'Random Forest Classifier: ':RandomForestClassifier(),
            'MLP Neural Network: ':MLPClassifier(max_iter=500),
            'Support Vector Classifier: ':SVC()
            }

    if StackML_model:
        models['StackML: '] = StackML_model

    errors = []
    print('\n', 'Model Accuracy Scores (1 is best):')

    for i, dataset in enumerate(datasets):

        Xtrain, ytrain, Xtest, ytest = get_benchmark_data(dataset)

        print('\n', f'Benchmark {i+1}:')
        print('--------------------------------')
        print(f'Dataset: {dataset}', '\n')

        for model_name, model in models.items():

            try:
                model.fit(Xtrain, ytrain)
                print(model_name, round(model.score(Xtest, ytest),4))

            except Exception as err:
                print(f'Failed to train {model_name}')
                errors.append(err)

    if errors:
        print('\n', 'Errors:')
        for error in errors:
            print(error)


def regression_benchmark(StackML_model=None, datasets = None, models=None):

    if not datasets:
        datasets = [
            '529_pollen',
            '622_fri_c2_1000_50',
            '649_fri_c0_500_5',
            '690_visualizing_galaxy',
            ]

    if not models:
        models = {
            'Linear Regression: ':LinearRegression(),
            'Random Forest Regressor: ':RandomForestRegressor(),
            'MLP Neural Network: ':MLPRegressor(max_iter=500),
            'Support Vector Regressor: ':SVR()
            }


    if StackML_model:
        models['StackML: '] = StackML_model

    errors = []
    print('\n', 'Model R2 Scores (1 is best):')


    for i, dataset in enumerate(datasets):

        Xtrain, ytrain, Xtest, ytest = get_benchmark_data(dataset)

        print('\n', f'Benchmark {i+1}:')
        print('--------------------------------')
        print(f'Dataset: {dataset}', '\n')

        for model_name, model in models.items():

            try:
                model.fit(Xtrain, ytrain)
                print(model_name, round(model.score(Xtest, ytest),4))

            except Exception as err:
                print(f'Failed to train {model_name}')
                errors.append(err)

    if errors:
        print('\n', 'Errors:')
        for error in errors:
            print(error)


def test_all_regression(model):
    pass

def test_all_classification(model):
    pass


def quick_test():
    model = s.StackML(verbose=True)
    datasets = [
        '529_pollen',
        '622_fri_c2_1000_50',
        # '690_visualizing_galaxy',
        '649_fri_c0_500_5',
        ]

    for i, dataset in enumerate(datasets):
        Xtrain, ytrain, Xtest, ytest = get_benchmark_data(dataset)
        model.fit(Xtrain, ytrain)
        print(f'{dataset}:', round(model.score(Xtest, ytest),4))
        model.plot_prediction(filename=dataset)


if __name__ == '__main__':
    quick_test()